1. Field of the Invention
The present invention relates generally to waveguide structures and more particularly to waveguide multiplexers/demultiplexers.
2. Description of the Related Art
Frequency-division multiplexing is the process of transmitting a plurality of input signals over a common transmission path by assigning a different frequency channel for each signal. Thus, the combined signals can subsequently be separated by filtering and by providing separate transmission paths for the filtered signals. The filtering and providing processes are those of frequency-division demultiplexing. Because demultiplexing is the inverse function of multiplexing, the following discussion is restricted, for simplicity, to multiplexing.
Low-loss, high-power frequency-division multiplexing in the microwave region is facilitated by the use of waveguide multiplexers which typically form a plurality of input ports for the reception of microwave input channel signals and a single output port for delivery of the multiplexed signals onto a common transmission path. Generally, this path leads to common signal-processing structures, e.g., a microwave amplifier or a radiating antenna.
In conventional waveguide multiplexers, a plurality of input waveguides (typically referred to as tee's) are joined to a single output waveguide (typically referred to as a manifold) in a way that enhances electromagnetic signal transfer. For example, each tee is arranged to form an E-plane junction with the manifold in one exemplary multiplexer structure and an H-plane junction in another. In most multiplexer waveguide structures, the manifold has an open-circuited end for transmission of the multiplexed input signals. Opposite the open-circuited end, the manifold has a short-circuited end and the tees are spaced by selected distances from the shorted end. Each tee also terminates in a short-circuited end and forms an aperture in this short-circuited end for signal access to that tee. A waveguide filter is coupled to the aperture so that channel filtering is associated with signal transmission through the tee.
In practice, a number of problems complicate multiplexer design. First, each input signal travels down its respective tee and splits into two signals which propagate in opposite directions along the manifold. One signal propagates towards the manifold's open-circuited end and the other propagates to, and is reflected from, the manifold's short-circuited end. Tee and manifold distances must therefore be carefully chosen so that each reflected signal from the manifold's short-circuited end adds to signals entering the manifold from that reflected signal's respective tee, i.e., these signals must be substantially in phase when they meet.
Secondly, the reflected signals from the manifold's short-circuited end again split as they successively reach each tee, with one signal portion propagating down the manifold and the other portion propagating up that tee and being reflected from that tee's short-circuited end. Tee and manifold distances must also be chosen so that signals reflected from tee shorted ends arrive in phase with signals entering the manifold from that reflected signal's respective tee.
Because they lie in different frequency channels, each of the input signals propagates with a different guide wavelength .lambda.g. A successful multiplexer design must therefore take the different propagation wavelengths into account and realize a dimensional layout that enhances signal additions at each tee so as to enhance the transmitted channel energy at the manifold's open-circuited end.
Multiplexer design is further complicated by impedance mismatches at the junctions of the tees and the manifold which generate additional signal reflections. An acceptable multiplexer design must reduce these impedance mismatches as much as possible and yet accommodate the reflected signals from the remaining mismatches.
Impedance mismatches can result in an apparent electrical short circuit wall at a specific frequency. A manifold resonance can be created between this apparent electrical wall and the manifold's short-circuited end or any one of the tee short-circuited ends. These types of resonances further degrade multiplexer performance.
In addition, multiplexer transmission-line discontinuities (e.g., tee-manifold junctions) generate higher-order electromagnetic modes. Because multiplexers are generally associated with nonlinear processes (e.g., high-power amplification), the input signals typically include frequency harmonics and this combination of discontinuities and harmonics generates higher-order harmonic modes which propagate in the multiplexer with different guide wavelengths. At other discontinuities (e.g., downstream waveguide junctions), energy is exchanged between these propagating modes. A successful multiplexer design must also control the energy exchanges of propagating higher-order modes in order to enhance the transmitted channel energy.
These complications of multiplexer design generally increase exponentially with each additional frequency channel that is included in the multiplexer. It has been found, for example, that although a satisfactory design can be found relatively quickly for an eight channel waveguide multiplexer, a satisfactory design for a sixteen channel multiplexer is exceedingly difficult to obtain.